Table 1.

R² of Naef and Magnasco model (Naef and Magnasco, 2003) (NM) and the position-dependent secondary-structure attenuated affinity model (PPSA) for the seven data sets used in this study. Results presented as average R² ± SD.

Data set	naa	npb		NM	PSAAc
Latin Square	42	248152	PM	0.17±0.009	0.184±0.010
		248152	MM	0.40±0.009	0.416±0.009
Choe (Choe et al., 2005)	6	195994	PM	0.20±0.022	0.216±0.025
		195994	MM	0.46±0.017	0.49±0.017
Leukemia (Armstrong et al., 2002)	72	201800	PM	0.49±0.063	0.51±0.062
		201800	MM	0.60±0.036	0.61±0.035
Etoposide response (Fodor et al., 2006)	60	496468	PM	0.05±0.040	0.06±0.040
		496468	MM	0.11±0.062	0.12±0.062
BK knockout (Meredith et al., 2006;Pyott et al., 2006)	20	496468	PM	0.09±0.035	0.10±0.036
		496468	MM	0.29±0.050	0.30±0.049
Splicing microarray (Sugnet et al., 2006)	75	505916	PM	0.30±0.062	0.31±0.063
Malaria (Le Roch et al., 2003)	17	173262	PM	0.36±0.043	0.38±0.043

^ana: number of chips.

^bnp: number of probes.

^cThe differences in R² between NM and PSAA are all statistically significant (P < 10^-3) using paired one-sided Wilcoxon test.